Watching the Online Safety Bill’s progress through Parliament is like riding the worst kind of roller coaster: it makes you feel queasy, everyone else seems to be having fun, and just when you think it cannot get worse you are flipped upside down and put into free fall, using your remaining strength to keep your lunch where it should be. Any hope that the government would address the numerous problems embedded in the Bill is evaporating quickly. Whilst its supporters have praised the Bill as a cutting edge child safety proposal, a diverse range of policy analysts, industry representatives, civil rights groups, academics, intelligence experts, lawyers and others have warned for years that the Bill represents a grave threat to free speech, privacy and security. The government has not heeded these warnings. Instead, it has recently suggested changes to the deeply flawed Bill that will make it worse. As a result, the UK is poised to have the most restrictive internet speech laws in the club of Western liberal democracies.

WhatsApp and Wikipedia may shortly be unavailable

The Bill is extensive (255 pages), and it has been on a long journey from a 2017 green paper to its current form, which the House of Lords continues to debate. It is ambitious, aiming to deliver on a manifesto commitment to make the UK “the safest place in the world to be online”. The Bill imposes obligations on search engines and other online platforms, and it backs them with eye watering fines of up to 10 per cent of global annual turnover or £18 million (whichever is greater).

Amongst the obligations included in the Bill that the Commons sent to the Lords are those associated with illegal content, and content that is legal but harmful to children. The Bill also imposes requirements on websites that host pornographic content. It allows Ofcom to require platforms to use “accredited technology” to search private messages for child sexual abuse material.

The Bill includes other provisions, such as those that target fraudulent advertising, but it is the parts associated with online speech, private messages and pornographic content that have been criticised the most. As I have explained before, these provisions provide incentives for online platforms to remove troves of legal and valuable speech in an abundance of caution, also putting end-to-end encrypted communications at risk.

It is therefore not surprising that WhatsApp, Signal and Wikipedia have each committed to refusing to weaken their users’ privacy or compromise on their commitments to only collect small amounts of data. As a result, each service may be unavailable to British users shortly after the Online Safety Bill passes. Unfortunately, rather than amend the Bill to provide reassurance to the tens of thousands of services that will be affected by its requirements and even more users who rely on them, the government has introduced amendments that make the Bill worse.

One of the most significant proposed changes to the Bill toughens requirements for services that host or allow pornography and other content considered especially harmful to children (“Primary priority content that is harmful to children”). The current version of the Bill does not provide a list of what constitutes such content; it merely identifies age verification as an option for limiting access to pornography. Under the government’s proposed changes, services will have to use “highly effective” age verification or age assurance techniques to prevent children from accessing pornography and “primary priority content that is harmful to children”, if such content is not already prohibited by terms of service. Such content includes material associated with suicide, self harm and eating disorders. Age verification and assurance measures vary, but one possible measure is the use of facial scanning.

Online services will remove legal and unobjectionable content en masse

Although well-intended, these amendments will result in less valuable and legal content for adults as well as children. As Wikipedia, not a site with a reputation as a pornographic content provider, noted a few weeks ago that educational content related to sexuality could be misinterpreted as pornography. Although authors of the government amendment may have pornographic websites in mind, there is no doubt that many other services that do not market themselves as providers of pornographic content will feel the effects of the government’s amendments.

The proposed age assurance and verification requirements will not only normalise persistent and creepy surveillance, but they will also compound one of the worst likely effects of the Bill: online services removing content in an abundance of caution. For services that host millions of pieces of user-generated content a day, it is not feasible for content moderation policies to account for the nuances that are often involved in determining whether a piece of content promotes suicide, or provides resources for those seeking help for suicidal ideation; or whether a sex education video is pornography. They will instead remove legal and unobjectionable content en masse from their platforms to stay on the safe side.

It is a notable sign of political change that a Conservative government is pressing ahead with the Online Safety Bill. Many of today’s Conservative MPs and peers seem to have abandoned faith in the non-government institutions between the individual and the state. Families, charities, religious institutions, the media, educational organisations and technology companies that empower parents with surveillance and blocking techniques have been apparently ignored in favour of the blunt power of legislation. Unfortunately, it is likely that the Online Safety Bill will pass. In the wake of this decision, the government may struggle to explain to the public why the internet is less open and online messaging less secure. Critics must ensure that the Bill’s supporters will not be able to say that they were not warned.